It's difficult to get the science right, because controls are difficult to come by. In TFA, he refers to two important longitudinal studies which indicate that

It's difficult to get the science right, because controls are difficult to come by. In TFA, he refers to two important longitudinal studies which indicate that there is a persistent positive effect between early readers and non-early readers. I have not read those studies, but I can't help but wonder how well they controlled for confounding factors. That is, that early reading is an effect or rather evidence, of their long run aptitude in school. That if the early readers were compared not just with the general population, but with students who were not early readers but had the same income levels, parental education level, family structure, etc. He also claims without evidence that any child can be taught to read at an early age. But it sounds like in his sample N=2. If there is evidence where a random sample (rather than a self selecting sample, which would introduce bias) of children are taught reading early and have a persistent benefit I would like to see it. That being said, good for him,and good for him to recommend this to others. It seems clear that his children seem happy, well-adjusted and achieving their academic potential. Given his personal experience, why wouldn't he share it? But I don't think the evidence is there to claim this is universal.